Research On Particle Size Detection Method Of Corn Pulverization Based On Machine Vision

In the Agricultural and industrial production process,the appropriate corn crushed particle size can improve the energy utilization of corn,which is of great practical significance to improve the health of livestock and poultry animals,improve their production performance,and promote the healthy and rapid development of the Chinese livestock industry.For the traditional maize crushed particle size detection method is single,inefficient and not high precision,the study took the red single 208 maize particles which were commonly planted in Hohhot area of Inner Mongolia as the study object,the maize crushed particle size was efficiently and precisely detected based on the improved watershed algorithm by acquiring maize crushed particle images through machine vision system.In addition,an intelligent detection system of corn crushed particle size was developed based on the App Designer module of Matlab.Details of the study and results are as follows:1.The study used a CPS-420 hammer mill for crushed maize particles.Different particle sizes were obtained by adjusting the crusher speed and replacing the sieves with different apertures,and then the crushed maize particles were graded by manual sieving.2.Comparison and analysis of different preprocessing algorithms,the algorithms of grayscale conversion,median filtering and Canny edge detection with threshold segmentation were selected to preprocess the image of crushed corn particles.Besides,the improved watershed algorithm was used to segment the image edges.Compared with the traditional watershed algorithm,the improved watershed algorithm solves the over-segmentation problem of the traditional watershed algorithm and separates the corn crushed particle images more accurately,which eventually achieves a more desirable edge segmentation of the adhered particles.3.Based on the relationship between the actual area and pixel area,the scale factor of the imaging system was derived using the pixel area occupied by known calibrators under the imaging system.Then the actual characteristic parameters in the image of crushed corn particles were calculated.Further,the segmentation area was calibrated,and the distribution of the crushed corn particles was counted.The feasibility of machine vision was verified by manual sieving tests,and error analysis and particle size distribution mapping were performed based on the detection results.It is shown that the difference between the detected particle size distribution results and the actual sieving results is minor,and the cumulative error rate for identifying different levels of mixed particles at different levels is less than 6.73%.4.A human-computer interaction interface was developed based on the Matlab App Designer module,which was combined with a written image processing algorithm for corn crush size detection to develop a Maize particle size intelligent detection system.The system mainly includes an image acquisition region,image parameter adjustment region,image processing region and result saving region,image loading,image denoising,image binarization,image segmentation,particle size detection and particle size data storage by using the information transfer with Matlab and Excel.This study verifies the feasibility of machine vision for corn crushed particle size detection,and provides a new idea for non-destructive detection of forage particle size,while laying a theoretical foundation for the intelligent development of feed particle size detection.